Quality control
2022-08-12
Image and cell-level quality control
The following section discusses possible quality indicators for data obtained by IMC and other highly multiplexed imaging technologies. Due to the complexity of the data, the quality metric range across the image and single-cell levels and the chapter is sectioned as such.
Read in the data
We will first read in the data processed in previous sections. As
compensation on all images required too much memory (>45GB), we are
not using the compensated images below. For convenience, we will re-set
the channelNames to their biological targets.
path <- "/mnt/Multimodal_Imaging_Daria/_Data_Analysis/_tmp_daria/Image_analysis/20220723_Steinbock_Mesmer_IFbased"
images <- readRDS(file.path(path,"data/images.rds"))
masks <- readRDS(file.path(path,"data/masks.rds"))
spe <- readRDS(file.path(path,"data/spe.rds"))
channelNames(images) <- rownames(spe)Segmentation quality control
The first step after image segmentation is to observe its accuracy.
Without having ground-truth data readily available, a common approach to
segmentation quality control is to overlay segmentation masks on
composite images displaying channels that were used for segmentation.
The cytomapper
Bioconductor package supports exactly this tasks by using the
plotPixels function.
Here, we select 3 random images and perform image- and channel-wise normalization (channels are first min-max normalized and scaled to a range of 0-1 before clipping the maximum intensity to 0.2).
set.seed(20220118)
img_ids <- sample(seq_len(length(images)), 3)
# Normalize and clip images
cur_images <- images[img_ids]
cur_images <- normalize(cur_images, separateImages = TRUE)
cur_images <- normalize(cur_images, inputRange = c(0, 0.2))
plotPixels(cur_images,
mask = masks[img_ids],
img_id = "image_id",
missing_colour = "white",
colour_by = c("CD8a", "CD3", "CD4", "GD2", "DAPI"),
colour = list(CD8a = c("black", "yellow"),
CD3 = c("black", "red"),
CD4 = c("black", "green"),
GD2 = c("black", "cyan"),
DAPI = c("black", "blue")),
image_title = NULL,
legend = list(colour_by.title.cex = 0.7,
colour_by.labels.cex = 0.7))
To zoom into the image you can right click and select
Open Image in New Tab. We can see that nuclei are centered
within the segmentation masks and all cell types are correctly segmented
A common challenge here is to segment large (e.g. GD2 - in cyan)
versus small (e.g. T cells - in red). However, the segmentation
approach here appears to correctly segment cells across different
sizes.
An additional approach to observe cell segmentation quality and potentially also antibody specificity issues is to visualize single-cell expression in form of a heatmap. Here, we sub-sample the dataset to 2000 cells for visualization purposes and overlay the sample from which the cells were extracted.
library(RColorBrewer)
n <- length(unique(spe$sample))
qual_col_pals = brewer.pal.info[brewer.pal.info$category == 'qual',]
col_vector = unlist(mapply(brewer.pal, qual_col_pals$maxcolors, rownames(qual_col_pals)))
metadata(spe)$color_vectors$sample <- setNames(sample(col_vector, n),
unique(spe$sample))library(dittoSeq)
library(viridis)
cur_cells <- sample(seq_len(ncol(spe)), 2000)
dittoHeatmap(spe[,cur_cells], genes = rownames(spe)[rowData(spe)$use_channel],
assay = "exprs", cluster_cols = TRUE, scale = "none",
heatmap.colors = viridis(100), annot.by = "sample",
annotation_colors = list(sample = metadata(spe)$color_vectors$sample))
dittoHeatmap(spe[,cur_cells], genes = rownames(spe)[rowData(spe)$use_channel],
assay = "exprs", cluster_cols = TRUE, scaled.to.max="TRUE",
heatmap.colors = viridis(100), annot.by = "sample",
annotation_colors = list(sample = metadata(spe)$color_vectors$sample))
We can differentiate between B-cells (CD20) and T cells (CD3). Some of the markers are specifically detected (e.g., Ki67, CD20, GD2) and others are more broadly detected (e.g. CD45, HLA-ABC, MPO).
Image-level quality control
Image-level quality control is often performed using tools that offer a graphical user interface such as QuPath and FIJI. Viewers that were specifically developed for IMC data can be seen here. In this section, we will specificaly focus on quantitative metrics to assess image quality.
It is often of interest to calculate the signal-to-noise ratio (SNR) for individual channels and markers. Here, we define the SNR as:
\[SNR = I_s/I_n\]
where \(I_s\) is the intensity of
the signal (mean intensity of pixels with true signal) and \(I_n\) is the intensity of the noise (mean
intensity of pixels containing noise). Finding a threshold that
separates pixels containing signal and pixels containing noise is not
trivial and different approaches can be chosen. Here, we use the
otsu thresholding approach to find pixels of the
“foreground” (i.e., signal) and “background” (i.e., noise). The SNR is
then defined as the mean intensity of foreground pixels divided by the
mean intensity of background pixels. We compute this measure as well as
mean signal intensity per image as well as the 95% confidence interval.
The plot below shows the
average SNR versus the average signal intensity across all
images.
library(tidyverse)
library(ggrepel)
library(EBImage)
image_subset <- getImages(images, 28:30)
cur_snr <- lapply(image_subset, function(img){
mat <- apply(img, 3, function(ch){
# Otsu threshold
thres <- otsu(ch, range = c(min(ch), max(ch)))
# Signal-to-noise ratio
snr <- mean(ch[ch > thres]) / mean(ch[ch <= thres])
# Signal intensity
ps <- mean(ch[ch > thres])
return(c(snr = snr, ps = ps))
})
t(mat) %>% as.data.frame() %>%
mutate(marker = colnames(mat)) %>%
pivot_longer(cols = c(snr, ps))
})
cur_snr <- do.call(rbind, cur_snr)
cur_snr %>%
group_by(marker, name) %>%
summarize(mean = mean(value),
ci = qnorm(0.975)*sd(value)/sqrt(n())) %>%
pivot_wider(names_from = name, values_from = c(mean, ci)) %>%
ggplot() +
# geom_errorbar(aes(y = log2(mean_snr), xmin = log2(mean_ps - ci_ps),
# xmax = log2(mean_ps + ci_ps))) +
# geom_errorbar(aes(x = log2(mean_ps), ymin = log2(mean_snr - ci_snr),
# ymax = log2(mean_snr + ci_snr))) +
geom_point(aes(log2(mean_ps), log2(mean_snr))) +
geom_label_repel(aes(log2(mean_ps), log2(mean_snr), label = marker)) +
theme_minimal(base_size = 15) + ylab("Signal-to-noise ratio") +
xlab("Signal intensity")
We observe ELAVL4, LUM and S100B to have high SNR but low signal intensity meaning that in general these markers are not abundantly expressed. The Iridium intercalator has high signal intensity but low SNR. This might be due to staining differences between individual nuclei where some nuclei are considered as background. We do however observe high SNR and sufficient signal intensity for the majority of markers. Results are not representative as only images [28:30] from sample 15-1320 were chosen due to memory issues.
Another quality indicator is the image area covered by cells (or
biological tissue). This metric identifies ROIs where little cells are
present, possibly hinting at incorrect selection of the ROI. We can
compute the percentage of covered image area using the metadata
contained in the SpatialExperiment object:
width_px <- 700
height_px <- 700
colData(spe) %>%
as.data.frame() %>%
group_by(sample_id) %>%
summarize(cell_area = sum(area),
no_pixels = mean(width_px) * mean(height_px)) %>%
mutate(covered_area = cell_area / no_pixels) %>%
ggplot() +
geom_point(aes(sample_id, covered_area)) +
theme_minimal(base_size = 15) +
ylim(c(0, 1)) +
theme(axis.text.x = element_text(angle = 90, hjust = 1, size = 2)) +
ylab("% covered area") + xlab("")
We observe that the images with low/high cell coverage are indeed
from the samples with a low/high cell number. These two images can now
be visualized using cytomapper.
# Normalize and clip images
img1="20211230_13-4339_BM_ROI_004"
img2="20220204_17-0390_BM_ROI_001"
img3="20211222_03-0313_BM_ROI_005"
img4="20220105_16-3300_BM_ROI_001"
cur_images <- images[c(img1, img4)]
cur_images <- normalize(cur_images, separateImages = TRUE)
cur_images <- normalize(cur_images, inputRange = c(0, 0.2))
plotPixels(cur_images,
mask = masks[c(img1, img4)],
img_id = "image_id",
missing_colour = "white",
colour_by = c("CD8a", "CD3", "CD4", "GD2", "DAPI"),
colour = list(CD8a = c("black", "yellow"),
CD3 = c("black", "red"),
CD4 = c("black", "green"),
GD2 = c("black", "blue"),
DAPI = c("black", "blue")),
legend = list(colour_by.title.cex = 0.7,
colour_by.labels.cex = 0.7))
Some samples might be oversegmented due to a high number of neutrophils, which have irregular shapes.
Finally, it can be beneficial to visualize the mean marker expression
per image to identify images with outlying cell type compositions. This
check does not indicate image quality per se but can highlight
biological differences. Here, we will use the
aggregateAcrossCells function of the scuttle
package to compute the mean expression per image. For visualization
purposes, we again asinh transform the mean expression
values.
library(scuttle)
image_mean <- aggregateAcrossCells(spe,
ids = spe$sample_id,
statistics="mean",
use.assay.type = "counts")
assay(image_mean, "exprs") <- asinh(counts(image_mean))
image_mean%>%dittoHeatmap(genes = rownames(spe)[rowData(spe)$use_channel],
assay = "exprs", cluster_cols = TRUE, scale = "none",
heatmap.colors = viridis(100),
annot.by = c("sample", "staining_date"),
annotation_colors = list(sample = metadata(spe)$color_vectors$sample,
staining_batch = metadata(spe)$color_vectors$staining_date),
show_colnames = FALSE)
We observe extensive biological variation across the images. Some images contain a high fraction of tumor cells (GD2), T cells (CD3) or proliferating cells (Ki67). These differences will be further studied in the following chapters.
Cell-level quality control
In the following paragraphs we will look at different metrics and visualization approaches to assess data quality (as well as biological differences) on the single-cell level.
Related to the signal-to-noise ratio (SNR) calculated above on the pixel-level, a similar measure can be derived on the single-cell level. Here, we will use a two component Gaussian mixture model for each marker to find cells with positive and negative expression. The SNR is defined as:
\[SNR = I_s/I_n\]
where \(I_s\) is the intensity of
the signal (mean intensity of cells with positive signal) and \(I_n\) is the intensity of the noise (mean
intensity of cells lacking expression). We calculate the SNR and signal
intensity by fitting the mixture model across the transformed counts of
all cells contained in the SpatialExperiment object.
library(mclust)
set.seed(220224)
mat <- apply(assay(spe, "exprs"), 1, function(x){
cur_model <- Mclust(x, G = 2)
mean1 <- mean(x[cur_model$classification == 1])
mean2 <- mean(x[cur_model$classification == 2])
signal <- ifelse(mean1 > mean2, mean1, mean2)
noise <- ifelse(mean1 > mean2, mean2, mean1)
return(c(snr = signal/noise, ps = signal))
})
cur_snr <- t(mat) %>% as.data.frame() %>%
mutate(marker = colnames(mat))
cur_snr %>% ggplot() +
geom_point(aes(log2(ps), log2(snr))) +
geom_label_repel(aes(log2(ps), log2(snr), label = marker)) +
theme_minimal(base_size = 15) + ylab("Signal-to-noise ratio") +
xlab("Signal intensity")
This analysis is more representative as it considers all cells, instead of cells of just one sample. We see that the nuclear channels have the highest signal intensity, but quite low SNR. CD14, CD3, GD2, CD8a, GZMB, Ki-67 and CD4 are some of the best markers and also have high SNR and intermediate signal intensity here. CD45 and HLA-ABC are expressed on a lot of cells and have a high signal. Some markers make no sense, such as ELAVL4 and PNMT, they are not nearly as good as CD20. HLA-DR should not have such as bad SNR and should also not be expressed on some many cells (as seen in heatmap above). Maybe this is caused by transformation?
Next, we observe the distributions of cell size across the individual images. Differences in cell size distributions can indicate segmentation biases due to differences in cell density or can indicate biological differences due to cell type compositions (tumor cells tend to be larger than immune cells).
colData(spe) %>%
as.data.frame() %>%
group_by(sample_id) %>%
ggplot() +
geom_boxplot(aes(sample_id, area)) +
theme_minimal(base_size = 15) +
theme(axis.text.x = element_text(angle = 90, hjust = 1, size = 2)) +
ylab("Cell area") + xlab("")
summary(spe$area)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 5.00 36.00 51.00 55.85 69.00 748.00The median cell size is 51 pixels with a median major axis length of
9. The largest cell has an area of 748 pixels which relates to a
diameter of r round(sqrt(max(spe$area)), digits = 1) pixels
assuming a circular shape. Overall, the distribution of cell sizes is
similar across images with image Patient4_005 and
Patient4_007 showing reduces average cell size. These
images contain fewer tumor cells which can explain the smaller average
cell size.
We detect very small cells in the dataset and will remove them. The chosen threshold is arbitrary and needs to be adjusted per dataset.
sum(spe$area < 5)## [1] 0spe <- spe[,spe$area >= 5]Another quality indicator can be the number of cells per image divided by the image size. This measure relates to the percentage of image area covered by cells as explained above.
colData(spe) %>%
as.data.frame() %>%
group_by(sample_id) %>%
summarize(cell_count = n(),
no_pixels = mean(width_px) * mean(height_px)) %>%
mutate(cells_per_mm2 = cell_count/(no_pixels/1000000)) %>%
ggplot() +
geom_point(aes(sample_id, cells_per_mm2)) +
theme_minimal(base_size = 15) +
theme(axis.text.x = element_text(angle = 90, hjust = 1, size = 2)) +
ylab("Cells per mm2") + xlab("")
The number of cells per mm\(^2\) varies across images which also depends on the number of tumor/non-tumor cells, but also the number of neutrophils which are more prone to oversegmentation.
The data presented here were stained on individual slides. Also, samples were pre-processed in different locations. These (and other) technical aspects can induce staining differences between samples or batches of samples. Observing potential staining differences can be crucial to assess data quality. We will use ridgeline visualizations to check differences in staining patterns:
multi_dittoPlot(spe, vars = rownames(spe)[rowData(spe)$use_channel],
group.by = "year", plots = c("ridgeplot"),
assay = "exprs",
color.panel = metadata(spe)$color_vectors$year)
We observe variations in the distributions of marker expression across years. These variations arise mainly from sample preparation differences between samples. Some markers seem to be more problematic such as HLA-ABC, CD274, CD24, CD45 and MPO. HLA-ABC and CD274 might be caused by staining artifacts.
##Dimensionality reduction on arcsinh-transformed counts Finally, we
will use non-linear dimensionality reduction methods to project cells
from a high-dimensional (40) down to a low-dimensional (2) space. For
this the scater
package provides the runUMAP and runTSNE
function. To ensure reproducibility, we will need to set a seed; however
different seeds and different parameter settings (e.g. the
perplexity) parameter in the runTSNE function
need to be tested to avoid interpreting visualization artefacts. For
dimensionality reduction, we will use all channels that show biological
variation across the dataset. However, marker selection can be performed
with different biological questions in mind.
We will parallelize the tasks to speed up the computation. In order to reduce the required memory per CPU, we will delete unused variables.
library(scater)
rm(selected_patients, selected_roi, selecting_staining_dates, staining_batch, staining_date, test, width_px, year)## Warning in rm(selected_patients, selected_roi, selecting_staining_dates, :
## object 'selected_patients' not found## Warning in rm(selected_patients, selected_roi, selecting_staining_dates, :
## object 'selected_roi' not found## Warning in rm(selected_patients, selected_roi, selecting_staining_dates, :
## object 'selecting_staining_dates' not found## Warning in rm(selected_patients, selected_roi, selecting_staining_dates, :
## object 'staining_batch' not found## Warning in rm(selected_patients, selected_roi, selecting_staining_dates, :
## object 'staining_date' not found## Warning in rm(selected_patients, selected_roi, selecting_staining_dates, :
## object 'test' not found## Warning in rm(selected_patients, selected_roi, selecting_staining_dates, :
## object 'year' not foundrm(images)
gc() #free unused space## used (Mb) gc trigger (Mb) max used (Mb)
## Ncells 10946593 584.7 16738248 894.0 16738248 894.0
## Vcells 522025645 3982.8 8132970981 62049.7 7099076349 54161.7set.seed(220225)
spe <- runUMAP(spe, subset_row = rowData(spe)$use_channel, exprs_values = "exprs", BPPARAM = MulticoreParam())
spe <- runTSNE(spe, subset_row = rowData(spe)$use_channel, exprs_values = "exprs", BPPARAM = MulticoreParam()) After dimensionality reduction, the low-dimensional embeddings are
stored in the reducedDim slot.
reducedDims(spe)## List of length 4
## names(4): UMAP TSNE UMAP_raw UMAP_loghead(reducedDim(spe, "UMAP"))## [,1] [,2]
## 20211222_03-0313_BM_ROI_005_1 9.459925 -0.8084583
## 20211222_03-0313_BM_ROI_005_2 9.450486 -0.9311345
## 20211222_03-0313_BM_ROI_005_3 9.589887 -0.8861686
## 20211222_03-0313_BM_ROI_005_4 9.542805 -0.9048243
## 20211222_03-0313_BM_ROI_005_5 9.561148 -0.8776727
## 20211222_03-0313_BM_ROI_005_6 9.549628 -0.9011635Visualization of the low-dimensional embedding facilitates assessment
of potential “batch effects”. The dittoDimPlot function
allows flexible visualization. It returns ggplot objects
which can be further modified.
library(patchwork)
library(viridis)
p1 <- dittoDimPlot(spe, var = "year", reduction.use = "UMAP", size = 0.2, legend.show = FALSE) +
scale_color_manual(values = metadata(spe)$color_vectors$year) +
ggtitle("year on UMAP")
p2 <- dittoDimPlot(spe, var = "year", reduction.use = "TSNE", size = 0.2, legend.size=2) +
scale_color_manual(values = metadata(spe)$color_vectors$year) +
ggtitle("year on TSNE") +
theme(legend.text=element_text(size=5))
(p1 + p2) 
p3 <- dittoDimPlot(spe, var = "sample", reduction.use = "UMAP", size = 0.2, legend.show = FALSE) +
scale_color_manual(values = metadata(spe)$color_vectors$sample) +
ggtitle("sample on UMAP")
p4 <- dittoDimPlot(spe, var = "sample", reduction.use = "TSNE", size = 0.2, legend.size=2) +
scale_color_manual(values = metadata(spe)$color_vectors$sample) +
ggtitle("sample on TSNE") +
theme(legend.text=element_text(size=2))
(p3 + p4)
multi_dittoDimPlot(
spe,
vars = rownames(spe)[rowData(spe)$use_channel],
reduction.use="UMAP",
assay="exprs",
legend.size=0.2,
size=0.2,
ncol=3,
nrow=13,
min.color="gray90",
max.color="red"
) +
theme(legend.text=element_text(size=1))
## NULLWe observe a strong separation of T-cells (CD3+ cells) between the patients. However a seperation of tumor (GD2+, CD56+) and non-tumor (CD45+) cells cannot really be seen. Therefore we will also run the UMAP on the non-transformed (raw) counts.
##Dimensionality reduction on raw counts Now we will try to run the UMAP on the non-transformed raw counts to see the effect of arcsinh-transformation.
set.seed(220225)
spe <- runUMAP(spe, subset_row = rowData(spe)$use_channel, exprs_values = "counts", name="UMAP_raw", BPPARAM = MulticoreParam()) multi_dittoDimPlot(
spe,
vars = rownames(spe)[rowData(spe)$use_channel],
reduction.use="UMAP_raw",
assay="exprs",
legend.size=0.2,
size=0.2,
ncol=3,
nrow=13,
min.color="gray90",
max.color="red"
) +
theme(legend.text=element_text(size=1))
## NULLWe see that the result is much worse!
##Dimensionality reduction on log-transformed counts As a final comparison, we will run the UMAP on log-transformed counts.
set.seed(220225)
spe <- runUMAP(spe, subset_row = rowData(spe)$use_channel, exprs_values = "log", name="UMAP_log", BPPARAM = MulticoreParam()) multi_dittoDimPlot(
spe,
vars = rownames(spe)[rowData(spe)$use_channel],
reduction.use="UMAP_log",
assay="exprs",
legend.size=0.2,
size=0.2,
ncol=3,
nrow=13,
min.color="gray90",
max.color="red"
) +
theme(legend.text=element_text(size=1))
## NULLAs there is no improvement, we will stick with the arcsinh transformation.
Save objects
The modified SpatialExperiment object is saved for
further downstream analysis.
saveRDS(spe, file.path(path,"data/spe.rds"))Session Info
SessionInfo
## R version 4.2.0 (2022-04-22)
## Platform: x86_64-pc-linux-gnu (64-bit)
## Running under: Ubuntu 20.04.4 LTS
##
## Matrix products: default
## BLAS: /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so.3
## LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/liblapack.so.3
##
## locale:
## [1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C
## [3] LC_TIME=en_US.UTF-8 LC_COLLATE=en_US.UTF-8
## [5] LC_MONETARY=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8
## [7] LC_PAPER=en_US.UTF-8 LC_NAME=C
## [9] LC_ADDRESS=C LC_TELEPHONE=C
## [11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C
##
## attached base packages:
## [1] stats4 stats graphics grDevices utils datasets methods
## [8] base
##
## other attached packages:
## [1] patchwork_1.1.1 scater_1.24.0
## [3] mclust_5.4.10 scuttle_1.6.2
## [5] ggrepel_0.9.1 forcats_0.5.1
## [7] stringr_1.4.0 dplyr_1.0.9
## [9] purrr_0.3.4 readr_2.1.2
## [11] tidyr_1.2.0 tibble_3.1.8
## [13] tidyverse_1.3.2 viridis_0.6.2
## [15] viridisLite_0.4.0 dittoSeq_1.8.1
## [17] ggplot2_3.3.6 RColorBrewer_1.1-3
## [19] BiocParallel_1.30.3 cytomapper_1.9.1
## [21] SingleCellExperiment_1.18.0 SummarizedExperiment_1.26.1
## [23] Biobase_2.56.0 GenomicRanges_1.48.0
## [25] GenomeInfoDb_1.32.3 IRanges_2.30.0
## [27] S4Vectors_0.34.0 BiocGenerics_0.42.0
## [29] MatrixGenerics_1.8.1 matrixStats_0.62.0
## [31] EBImage_4.38.0
##
## loaded via a namespace (and not attached):
## [1] readxl_1.4.0 backports_1.4.1
## [3] systemfonts_1.0.4 plyr_1.8.7
## [5] sp_1.5-0 shinydashboard_0.7.2
## [7] digest_0.6.29 htmltools_0.5.3
## [9] magick_2.7.3 tiff_0.1-11
## [11] fansi_1.0.3 magrittr_2.0.3
## [13] ScaledMatrix_1.4.0 SpatialExperiment_1.6.1
## [15] googlesheets4_1.0.0 tzdb_0.3.0
## [17] limma_3.52.2 modelr_0.1.8
## [19] svgPanZoom_0.3.4 R.utils_2.12.0
## [21] svglite_2.1.0 jpeg_0.1-9
## [23] colorspace_2.0-3 rvest_1.0.2
## [25] haven_2.5.0 xfun_0.32
## [27] crayon_1.5.1 RCurl_1.98-1.8
## [29] jsonlite_1.8.0 glue_1.6.2
## [31] gtable_0.3.0 gargle_1.2.0
## [33] nnls_1.4 zlibbioc_1.42.0
## [35] XVector_0.36.0 DelayedArray_0.22.0
## [37] BiocSingular_1.12.0 DropletUtils_1.16.0
## [39] Rhdf5lib_1.18.2 HDF5Array_1.24.2
## [41] abind_1.4-5 scales_1.2.0
## [43] pheatmap_1.0.12 DBI_1.1.3
## [45] edgeR_3.38.4 Rcpp_1.0.9
## [47] xtable_1.8-4 dqrng_0.3.0
## [49] rsvd_1.0.5 httr_1.4.3
## [51] htmlwidgets_1.5.4 ellipsis_0.3.2
## [53] pkgconfig_2.0.3 R.methodsS3_1.8.2
## [55] farver_2.1.1 uwot_0.1.11
## [57] sass_0.4.2 dbplyr_2.2.1
## [59] locfit_1.5-9.6 utf8_1.2.2
## [61] labeling_0.4.2 tidyselect_1.1.2
## [63] rlang_1.0.4 later_1.3.0
## [65] munsell_0.5.0 cellranger_1.1.0
## [67] tools_4.2.0 cachem_1.0.6
## [69] cli_3.3.0 generics_0.1.3
## [71] broom_1.0.0 ggridges_0.5.3
## [73] evaluate_0.16 fastmap_1.1.0
## [75] fftwtools_0.9-11 yaml_2.3.5
## [77] knitr_1.39 fs_1.5.2
## [79] sparseMatrixStats_1.8.0 mime_0.12
## [81] R.oo_1.25.0 xml2_1.3.3
## [83] BiocStyle_2.24.0 compiler_4.2.0
## [85] rstudioapi_0.13 beeswarm_0.4.0
## [87] png_0.1-7 reprex_2.0.1
## [89] bslib_0.4.0 stringi_1.7.8
## [91] highr_0.9 RSpectra_0.16-1
## [93] lattice_0.20-45 Matrix_1.4-1
## [95] vctrs_0.4.1 pillar_1.8.0
## [97] lifecycle_1.0.1 rhdf5filters_1.8.0
## [99] BiocManager_1.30.18 jquerylib_0.1.4
## [101] RcppAnnoy_0.0.19 BiocNeighbors_1.14.0
## [103] irlba_2.3.5 cowplot_1.1.1
## [105] bitops_1.0-7 raster_3.5-21
## [107] httpuv_1.6.5 R6_2.5.1
## [109] promises_1.2.0.1 gridExtra_2.3
## [111] vipor_0.4.5 codetools_0.2-18
## [113] assertthat_0.2.1 rhdf5_2.40.0
## [115] rjson_0.2.21 withr_2.5.0
## [117] GenomeInfoDbData_1.2.8 hms_1.1.1
## [119] parallel_4.2.0 terra_1.6-7
## [121] grid_4.2.0 beachmat_2.12.0
## [123] rmarkdown_2.14 DelayedMatrixStats_1.18.0
## [125] googledrive_2.0.0 Rtsne_0.16
## [127] lubridate_1.8.0 shiny_1.7.2
## [129] ggbeeswarm_0.6.0